1
UNIVERSIDAD AUTÓNOMA DE MADRID Facultad de Medicina
Departamento de Pediatría
The effects of age, sex and genetics on the metabolic response to high fat diet intake in mice
Memoria para optar al grado de Doctor con Mención Internacional de la licenciada
Doña ALEJANDRA FREIRE FERNÁNDEZ-REGATILLO
DIRECTORES:
Prof. Dr. D. Jesús Argente Oliver Dra. Doña. Julie Ann Chowen King
Madrid, 2019
2
Este trabajo ha sido realizado con cargo a proyectos del Ministerio de Ciencia e Innovación (BFU2011-27492 and BFU2014-51836-C2-2-R), del Fondo de Investigación Sanitaria (PI1007047), y del CIBER de Fisiopatología de la Obesidad y Nutrición (CB06-03). Igualmente, ha sido financiado gracias a la Fundación de Endocrinología y Nutrición del Hospital Infantil Universitario Niño Jesús.
Alejandra Freire Fernández-Regatillo ha recibido financiación del programa de Ayudas Predoctorales FPU del Ministerio de Educación y Ciencia (FPU13/09009).
3
Me gustaría agradecer a tantas personas poder terminar esta tesis que no sé ni por dónde empezar.
Bueno, sí se por dónde empezar. Por mis directores, los doctores Jesús Argente Oliver y Julie Ann Chowen. Gracias por confiar en mí y por dejarme desarrollar mi tesis con vosotros, por vuestra paciencia, vuestro apoyo y todo lo que me habéis enseñado.
Gracias al Equipo P, a nuestras tardes de nachos y margaritas, nuestros cafés de emergencia y todos esos ratos buenos que hemos pasado juntos, en los que me habéis hecho evolucionar.
Gracias a Santi por tener siempre un gesto o una palabra bonita y, por supuesto, una mano para ayudar.
Gracias a Laura, Sandra y Vicente por todo lo que he aprendido con vosotros también.
Gracias a las chicas y chicos que han pasado por el laboratorio, se han dejado enseñar y se han reído con nosotros. En especial a MJ y Lourdes, que cada una a su manera ha sido todo amor.
Gracias a mis amigos y amigas fuera del laboratorio, que me han ayudado a ver el mundo más allá.
Y no me quiero dejar fuera a Estefi, que me introdujo en la investigación y a quien considero mi inspiración. Y, con ella, a toda la gente del C01, con especial mención a Luis Miguel y con mención aparte a mis Biolocas, que le dan el punto divertido a la ciencia.
Y con un agradecimiento a mi hermano, que es el mejor del mundo, otro a mis padres, a mis tíos y, por qué no, también a mis gatos, dejo esta sección porque tengo que enviar esto a imprimir ya.
Os quiero.
4
5
SUMMARY
6
7
Obesity and its secondary complications continue to increase worldwide and constitute one of the most important health care problems in many societies. It is now clear that the underlying causes of obesity can vary and that not all individuals have similar propensities to become obese and/or to develop obesity- associated pathologies. Indeed, the development of treatment strategies and drugs to curtail this epidemic span a wide range of targets. However, much is yet to be learned regarding metabolic control and individual differences in this process.
The aim of this thesis was to analyze the effects of age, sex and genetic background on the response to a high fat diet in mice. Puberty is a period of development with multiple changes occurring both in the central nervous system and systemically. However, little is known regarding the specific response during puberty to poor nutrition. Hence, we analyzed the metabolic response of both male and female peripubertal mice to a short-term high fat diet protocol. On the other extreme, aging increases the overall propensity to weight gain and some diseases, including neurodegenerative diseases such as Alzheimer’s. Hence, we employed a genetic model of Alzheimer’s disease in mice, submitting male and female mice to a long-term high fat diet or low fat diet and analyzing their metabolic response.
Moreover, as astrocytes are involved in both metabolic control and neuroprotection, we analyzed the response of hypothalamic astrocytes from the hypothalamus of males and females to palmitic acid, a saturated fatty acid commonly found in our diet, and to amyloid-β, which is an important component in the development of Alzheimer’s disease.
The results reported here indicate that during the peripubertal/pubertal period mice are less prone to excess weight gain to short-term high fat intake compared to that reported at other developmental ages. Mice of both sexes increased their energy intake, but body weight was not affected. However, metabolic profiles were modified in a sex dependant manner with males being more affected than females.
When mice were exposed to a high fat diet during later adulthood, we found that females gained more weight and fat mass than males, which is in contrast to what is commonly reported for young adults.
8
Key words: Age; Sex; High fat diet; Astrocytes; Sex steroids; Alzeimer’s disease.
9
RESUMEN
10
La obesidad y sus complicaciones secundarias continúan aumentando en todo el mundo y constituyen uno de los mayores problemas de salud en nuestra sociedad. Ahora sabemos que sus causas son variadas y que cada individuo muestra diferente propensión a desarrollar obesidad y/o patologías asociadas a esta. Sin duda, el desarrollo de estrategias terapéuticas y medicamentos contra esta epidemia abarca un amplio rango de posibles dianas. Sin embargo, todavía queda mucho que aprender sobre el control del metabolismo y sus diferencias individuales.
El objetivo de esta tesis es analizar los efectos de la edad, el sexo y el trasfondo genético en la respuesta a una dieta alta en grasas en ratones. La pubertad es un período del desarrollo con múltiples cambios que ocurren tanto en el sistema nervioso central como a nivel periférico. Sin embargo, las respuestas específicas a una mala nutrición durante la pubertad no han sido estudiadas en profundidad. Por tanto, hemos analizado la respuesta metabólica de ratones de ambos sexos en edad peri-puberal, sometiéndoles a una dieta alta en grasa durante un corto período de tiempo. Por otra parte, el envejecimiento aumenta la propensión al sobrepeso y ciertas enfermedades, incluyendo trastornos neurodegenerativos como el de Alzheimer. Por ello, hemos utilizado un modelo genético de enfermedad de Alzheimer en ratones, sometiendo a machos y a hembras a una dieta alta o baja en grasas durante un período largo de tiempo para analizar su respuesta metabólica. Además, dado que los astrocitos están implicados tanto en el control metabólico como en la neuroprotección, hemos analizado la respuesta de astrocitos hipotalámicos de machos y hembras al ácido palmítico, un ácido graso saturado que se encuentra comúnmente en nuestra dieta, y al amiloide-β, un compuesto importante en el desarrollo de la enfermedad de Alzheimer.
Los resultados que aquí se exponen indican que los ratones en edad puberal/peripuberal son menos propensos a la ganancia de peso tras un período corto de dieta alta en grasas, en comparación con lo que se ha observado en otros estadios del desarrollo. Los ratones de ambos sexos aumentaron su ingesta calórica sin aumentar de peso. Sin embargo, los perfiles metabólicos se vieron
11
alterados de diferente manera en función del sexo, con un mayor efecto en machos que en hembras.
Cuando ratones de mediana edad fueron expuestos a una dieta alta en grasas encontramos que las hembras ganaron más peso y grasa que los machos, lo que contrasta con lo que se suele observar en adultos jóvenes.
Palabras clave: Edad; Sexo; Dieta alta en grasas; Astrocitos; Esteroides sexuales;
Enfermedad de Alzheimer.
12
13
INDEX
14
15
o Summary – page 5 o Resumen – page 9 o Introduction – page 17 o Hypothesis – page 35 o Objectives – page 39
o Material and Methods – page 43 o Results – page 59
o Discussion – page 89 o Conclusions – page 105 o Conclusiones – page 109 o References – page 113
16
17
INTRODUCTION
18
19 1. Obesity and high-fat diet intake
Overweight and obesity are important health problems in our era, as being obese has been associated with the development of type 2 diabetes, dyslipidemia, asthma, cardiovascular problems (Anteneh et al 2015, Martin-Rodriguez et al 2015) and even certain types of cancer (Hursting & Dunlap 2012) and neurological diseases like dementia (Emmerzaal et al 2015). The World Health Organization (WHO) defines obesity and overweight in adults by means of body mass index (BMI). The BMI of an individual is calculated as their body weight (in kilograms) divided by the square of their height (in meters). In adults, values over 25 kg/m2 are considered overweight and values over 30 kg/m2 are considered obesity.
Since 1975, obesity has tripled worldwide. In 2016, 13 % of adults and 7% of children and adolescents were obese, according to the WHO. This organization places much of the blame on an increase in high-fat diet intake and a decrease of physical activity. That is, an imbalance between the intake and expenditure of energy (Garrow 1988). These changes are often a consequence of multiple environmental and social factors, like urban development, sedentarism, working habits, food processing, education and a lack of social politics in these sectors (Who 2015).
2. Neuroendocrine control of metabolism
The hypothalamus (Cone et al 2001, Kim et al 2014a, Schneeberger et al 2014), the brain area in charge of homeostatic control, is the principal center in the central nervous system (CNS) for integration of nutritional and hormonal inputs. Here, specialized neuronal circuits process this information and send signals to higher brain regions, resulting in the modification of appetite and energy expenditure (Webber et al 2015).
2.1. Neuronal circuits
Within the hypothalamus, the arcuate nucleus contains a specialized neuronal circuit known as the melanocortin system, the major network in metabolic regulation (Cone 2006). This system is formed fundamentally by two neuronal populations exerting opposite actions. One of them is orexigenic and is characterized by the co- expression of neuropeptide Y (NPY), Agouti-related protein (AgRP) and gamma-amino
20
butyric acid (GABA) (Hahn et al 1998, Krashes et al 2013) (Figure 1). The other is mainly anorexigenic and expresses pro-opiomelanocortin (POMC) derived peptides, particularly 𝛼-melanocyte stimulating hormone (𝛼-MSH), with a portion of these neurons co-expressing cocaine-and-amphetamine-regulated transcript (CART) (Balthasar et al 2004, Cowley et al 2003, Elias et al 1998). These POMC-neurons are also able to produce an orexigenic signal under certain circumstances, through alternate processing of POMC resulting in the production and release of 𝛽-endorphin (Dutia et al 2012, Grossman et al 2003, Koch et al 2015).
Hypothalamic POMC/CART neurons receive tonic inhibition from NPY/AgRP/GABA neurons (Horvath et al 1992) and excitatory signals from the ventromedial nucleus of the hypothalamus (VMH) (Sternson et al 2005). In addition, glutamatergic signals from within the arcuate nucleus stimulate both of these populations of neurons (Kiss et al 2005). Hormones such as leptin and insulin stimulate this anorexigenic neuronal population while inhibiting the orexigenic population (Horvath 2005, Schwartz et al 2000). On the other hand, ghrelin inhibits POMC/CART neurons and stimulates NPY/AgRP/GABA neurons (Cowley et al 2003, Chen et al 2017). Thus, these two important populations of neurons integrate the metabolic signals received from the circulation to determine the balance of energy need/energy excess.
The main targets of these neurons are cells expressing the melanocortin 4 receptor (MC4R) and the melanocortin 3 receptor (MC3R), which receive excitatory signals (e.g., 𝛼-MSH) from POMC/CART neurons. In contrast, AgRP released from NPY/AgRP/GABA neurons acts as an inverse agonist on these same receptors, resulting in inhibitory signals and opposing effects on energy expenditure (Fan et al 1997, Ollmann et al 1997).
The actions of these neurons constitute the first regulatory response of the CNS to the energy status of the body. These neurons send projections to other hypothalamic nuclei, such as the paraventricular nucleus (PVN), dorsomedial hypothalamus (DMH) and the lateral hypothalamic area (LHA) (Elias et al 1998, Elmquist et al 1998) where, in conjunction from direct inputs of both hormones and nutrients, will determine the output of these brain areas. The signals are then further processed by other brain areas, like the solitary nucleus, amygdala and prefrontal cortex (Abizaid & Horvath 2008, Horvath 2005, Kim et al 2014a, Schneeberger et al 2014, Schwartz et al 2000).
Together, these signals determine appetite and energy output.
21
Figure 1. The two main neuronal populations forming the melanocortin system and their interactions.
2.2. Role of astrocytes and other non-neuronal cells in energy homeostasis Neurons are not the only cell type involved in the control of metabolism within the CNS. To begin with, the epithelial cells found in the choroid plexus and lining the ventricles secrete cerebrospinal fluid (CSF) and factors involved in neurogenesis and development (Lehtinen et al 2011, Marques et al 2011, Parada et al 2008, Thouvenot et al 2006). They also exert barrier functions and express transporters for metabolites and receptors for sex steroids and leptin (Alves et al 2009, Hong-Goka & Chang 2004, Mitchell et al 2009, Quadros et al 2007, Saunders et al 2015, Spector 1989, Spector &
Johanson 2006). Together with tanycytes, they form the hypothalamic neurogenic niche (Mirzadeh et al 2008).
Tanycytes are polarized cells lining the walls of the third ventricle. They have a long process projecting into the hypothalamus or the median eminence (Rodriguez et al 2005). These specialized cells are involved in the control of the substance entry into the hypothalamus, as they express specific transporters and modulate the permeability of
22
the blood-CSF barrier (Langlet 2014, Langlet et al 2013a, Langlet et al 2013b).
Moreover, tanycytes participate in glutamate recycling, nutrient sensing and the conversion of thyroid hormones from its inactive to active form (Barrett et al 2007, Bolborea & Dale 2013, Frayling et al 2011, Nilaweera et al 2011). They are interconnected by gap junctions, which allow them to transmit coordinated signals through calcium waves (Orellana et al 2012).
Distributed throughout the CNS are microglial cells, the brain’s immune system.
These cells respond to brain damage, toxins and harmful conditions by changing their phenotype to a reactive form, producing cytokines, nitric oxide or reactive oxygen species (Dalmau et al 1997, Delgado et al 1998, Ginhoux et al 2013, Gonçalves et al 2012, Ling et al 1980). In addition, they regularly phagocyte cellular debris and other wastes and release gliotransmitters and metabolic factors (Aloisi 2001, Elkabes et al 1996, Gertig & Hanisch 2014). They are also involved in synaptic pruning, both during development and, at later ages, in response to specific signals (Batchelor et al 1999, Batchelor et al 2002, Zhong et al 2010).
Forming the blood-brain barrier (BBB), we find endothelial cells, pericytes and astrocytes. Endothelial cells strictly control the entry of substances from the blood to the CNS and pericytes, contractile cells surrounding the vessels, are involved in the regulation of blood flow and participate in the development and maintenance of the BBB, in addition to performing macrophage-like functions and being intricately involved in neuroinflammation (Abbott et al 2010).
Astrocytes are the most abundant cell type in the CNS. They exert multiple functions, beginning with their indispensable role in supplying physical and metabolic support to neurons. Astrocytes are also an important part of the BBB, transporting nutrients and metabolic factors and, thus, controlling what enters the CNS (Abbott et al 2006, Pellerin & Magistretti 1994). Also, they have a role in the maintenance of the barrier and the blood flow at this level (Janzer & Raff 1987, MacVicar & Newman 2015, Zonta et al 2003). These cells form a syncytium, as they are connected through gap junctions that enable the transport of some molecules -in a regulated fashion (Söhl
& Willecke 2004, Theis et al 2005)- and the transmission of calcium waves (Scemes &
Giaume 2006). Astrocytes store energy through glycogenesis (Cataldo & Broadwell 1986). They form part of the “tripartite synapse”, as they express glutamate transporters GLT1 and GLAST and take-up glutamate from the synaptic cleft (Pines et al 1992, Schmitt et al 1997), terminating transmission and preventing excitotoxicity (Araque et
23
al 1999, Ransom et al 2003). Astrocytes also regulate synaptogenesis and neuronal proliferation and differentiation during development (Clarke & Barres 2013, Nedergaard et al 2003). Importantly, they respond to injuries, foreign substances and infections by producing cytokines (Aschner 1998a, Aschner 1998b). In addition, astrocytes are important in glucose transport (Morgello et al 1995) and they can change their morphology, modifying synaptic structure/inputs (Clarke & Barres 2013, Nedergaard et al 2003).
2.3. Astrocytes in metabolism and nutrient sensing
Astrocytes express glutamate transporter GLUT2, sodium glucose transporter 1 (SGLT1), glucokinase (GCK) and KATP channels, all involved in the glucose sensing process (Leloup et al 2016, Steinbusch et al 2015). Glucose enters the cell through GLUT2 and it is phosphorylated by GCK, unless it undergoes glycogenesis. Then, glucose-6-phosphate produces pyruvate, which is converted in lactate via lactate dehydrogenase. Finally, lactate exits the cells through monocarboxylate transporter (MCT) 4 or 1 and it is taken up by neurons through MCT-2 (Elizondo-Vega et al 2015).
This process has been named the “astrocyte-neuron lactate shuttle” by Pellerin and Magistretti (Pellerin & Magistretti 1994). Lactate release by astrocytes also serves as a signal of energy availability to glucose-sensing neurons (Leloup et al 2016). Moreover, in response to an increase in glucose or to other factors, like neurotransmitters, astrocytes release endozepines, peptides that exert an anorexigenic effect on hypothalamic neurons (Lanfray et al 2013, Tonon et al 2013) and also have a role in unsaturated long-chain fatty acid metabolism in astrocytes (Bouyakdan et al 2015). The ability of astrocytes to take up glucose is regulated by various signals: it is increased by leptin (Fuente-Martin et al 2012), reduced by ghrelin (Fuente-Martin et al 2016) and it has been also demonstrated that insulin is important for glucose uptake by astrocytes (Garcia-Caceres et al 2016). Glucose storage in the form of glycogen is another important role of astrocytes on the management of energy availability. If needed, i.e., when there is an increase in neuronal activity (Cruz & Dienel 2002), astrocytic glycogen is processed to lactate via glycogenolysis. Glycogen production in astrocytes can be increased by diverse hormones, such as insulin, insulin-like growth factor (IGF)-1 and leptin (Bosier et al 2013, Heni et al 2011, Muhič et al 2015, Sartorius et al 2012). Also,
24
glycogenolysis has been shown to be promoted by ghrelin in hypothalamic astrocytes (Fuente-Martin et al 2016).
Astrocytes are the main metabolizers of lipids in the CNS. Moreover, they express proteins for lipid sensing, such as fatty acid translocase CD36 and peroxisome proliferator-activated receptor gamma (PPAR𝛾), a lipid-activated nuclear receptor that regulates transcription of several genes, some of them involved in lipid metabolism (Cristiano et al 2005, Heneka & Landreth 2007). Astrocytes are able to take up ketone bodies via MCT-1 and 4 (Bergersen et al 2002, Pierre & Pellerin 2005, Rafiki et al 2003) and also to synthesize them from fatty acids through fatty acid 𝛽-oxidation and release them to the extracellular space, where other cells use them as an energy source.
This process is important in nutrient sensing because ketone bodies produced and released by astrocytes are a sign for hypothalamic neurons of an excess of fatty acids, so they act to reduce food intake (after an initial HFD-induced hyperphagia) (Carneiro et al 2016, Le Foll et al 2014, Le Foll et al 2015). In fact, it has been reported that 𝛽- oxidation in hypothalamic astrocytes increases in obese mice fed a HFD and that tanycytes prevent the entry of an excess of saturated fatty acids into the hypothalamus (Hofmann et al 2016). Lipids can enter the CNS from the bloodstream as fatty acids, through simple diffusion or with the help of different fatty acid translocases, transport and binding proteins (Kamp et al 2003, Levin 1980), or as lipoproteins. The latter process is mediated by apolipoprotein E (ApoE), which is expressed in astrocytes and tanycytes and interacts with lipoprotein receptors (Boyles et al 1985). Thyroid hormones and leptin upregulate ApoE expression, thus regulating energy balance (Roman et al 2015, Shen et al 2009, Shen et al 2008). A summary of the main interactions of astrocytes in the hypothalamus is represented in Figure 2.
25
Figure 2. The complex functions of astrocytes on metabolism.
2.4. Hormones in metabolism 2.4.1. Leptin
Leptin is an anorexigenic hormone secreted principally by adipose tissue, although it can be produced also by the hypothalamus, stomach, liver, ovaries and placenta (Bado et al 1998, Hoggard et al 1997, Mantzoros 1999). When leptin reaches the hypothalamus, it inhibits NPY/AgRP/GABA neurons and stimulates POMC/CART neurons, thus decreasing food intake and increasing energy expenditure (Horvath 2005, Schwartz et al 2000, Simpson et al 2009). The leptin receptor (LepR) is also highly expressed in astrocytes (Hsuchou et al 2009) and leptin signaling in these cells is important for energy balance (Kim et al 2014b, Wang et al 2015).
Impairment of leptin expression or action leads to obesity, hyperphagia and low energy expenditure (Drel et al 2006, Farooqi et al 1999). Leptin levels in plasma correlate positively with BMI, as it is released to the bloodstream in proportion to the amount of adipose tissue (Argente et al 1997, Frederich et al 1995a), and also respond to the acute metabolic status, increasing with HFD intake and decreasing during fasting (Considine et al 1996, Friedman & Halaas 1998). In humans, leptin replacement
26
ameliorates the symptoms of obese leptin deficient patients, reducing adipose tissue (Farooqi et al 2002, Mantzoros et al 2011). However, obese individuals normally show elevated plasma leptin levels, implying the existence of a leptin resistance associated with weight gain (Frederich et al 1995b, Maffei et al 1995). It has been reported that HFD-induced obesity causes leptin resistance by two possible mechanisms: affecting the entry of leptin into the brain (Balland et al 2014) and reducing the central response to this hormone (Balland & Cowley 2015). Obesity, fasting and certain metabolic factors intervene in the regulation of leptin transport. Long term HFD induces leptin resistance in mice only when there are high levels of leptin in plasma (Knight et al 2010), suggesting that hyperleptinemia is one of the triggers for diet-induced leptin resistance. Also, hypothalamic inflammation linked to diet-induced obesity could be involved in leptin resistance by modifying the cellular and molecular systems that control energy homeostasis (de Git & Adan 2015). However, some studies suggest that the only response impaired in leptin resistance is the one from exogenous leptin, while the responsiveness to endogenous leptin remains intact (Flak & Myers 2016, Ottaway et al 2015).
Leptin participates in the control of peripheral glucose and lipid metabolism, promoting lipolysis and inhibiting lipogenesis in adipose tissue and liver (Frühbeck et al 1998, Hynes & Jones 2001, Kamohara et al 1997). Glucose and insulin favor the passage of leptin across the BBB (Kastin & Akerstrom 2001) and a high level of plasma triglycerides reduces it (Banks et al 2004). On the other hand, leptin decreases glucose- induced insulin secretion (Hynes & Jones 2001, Muzumdar et al 2003) and also affects the ability of hypothalamic astrocytes to transport glucose and glutamate (Fuente- Martin et al 2012).
2.4.2. Insulin
Insulin is a peptide mainly produced by the pancreas in response to high levels of glycemia. Insulin regulates glucose levels in blood by stimulating its uptake by other tissues and suppressing glucose productions by the liver (Kahn 1994). Insulin levels correlate positively with adiposity (Polonsky et al 1988). Insulin also acts as an anorexic hormone. The insulin receptor (IR) is expressed throughout the brain and is highly expressed in the arcuate nucleus (Van Houten et al 1979), where insulin activates POMC neurons and inhibits AgRP expression (Brüning et al 2000, Van Houten et al
27
1979), decreasing food intake and increasing energy expenditure (Qiu et al 2014, Woods et al 1979).
Astrocytic IRs have a role in the transport of insulin across the BBB, and insulin signaling in astrocytes is important in the regulation of glycemia (Garcia-Caceres et al 2016). Satiation hormones, like cholecystokinin (CCK) promote insulin transport into the brain and other hormones, like estradiol inhibit insulin functions in the brain, although not its transport (May et al 2016). The relationship and interactions between leptin and insulin need further research, as some studies have found that leptin enhances insulin sensitivity (Koch et al 2010) while others show that leptin inhibits insulin signaling in the brain (Sartorius et al 2012). In fact, these two hormones share some molecular pathways, although they have antagonic effects in some hypothalamic neurons (Xu et al 2005). Obesity-induced hyperinsulinemia leads to insulin resistance and type-2 diabetes (Moller & Flier 1991). Saturated fatty acids have been shown to be involved in insulin resistance in the hypothalamus (Benoit et al 2009) and in peripheral tissues (Boden et al 2002).
2.4.3. Ghrelin
Ghrelin is a gastrointestinal hormone that promotes food intake and is produced mainly in the stomach (Kojima et al 1999, Nakazato et al 2001). Ghrelin has also been shown to promote fat accumulation and control glucose homeostasis and energy expenditure (Müller et al 2015, Tschöp et al 2000, Wren et al 2001). During fasting, it is released and reaches the CNS, stimulating NPY/AgRP/GABA neurons and inhibiting POMC/CART neurons through its receptor, the growth hormone secretagogue receptor 1a or GHS-R1a, which is highly expressed in the hypothalamus (Kamegai et al 2001, Riediger et al 2003, Wang et al 2013, Zigman et al 2006). Ghrelin transport into the brain is enhanced by serum triglycerides, which are high during fasting, and is inhibited in obese individuals (Banks et al 2008). Also, diet-induced obesity can impair the hypothalamic response to ghrelin (Briggs et al 2010).
2.4.4. Sex steroids
Estrogens, androgens and progesterone are secreted principally by the reproductive organs (ovaries and testes), although they are also produced by other tissues such as adipose tissue (Simpson 2003) and the brain (Baulieu & Robel 1990).
28
The latter are called neurosteroids and can be synthesized from brain-borne cholesterol or from steroid precursors coming from the periphery, like pregnenolone, deoxycorticosterone and testosterone (Reddy 2010). The synthesis can take place in astrocytes, tanycytes, ependymal cells and oligodendrocytes (Jung-Testas & Baulieu 1998, Mensah-Nyagan et al 1999), and even in some neurons (Agís-Balboa et al 2006).
Sex steroids, mainly estrogens, are involved in the control of energy homeostasis at the hypothalamic level, reducing food intake (Butera & Czaja 1984, Czaja & Goy 1975, Gao et al 2007), increasing energy expenditure (Musatov et al 2007) and regulating the sensitivity to metabolic hormones like leptin, insulin and ghrelin (Clegg et al 2006, Clegg et al 2007). Their effect is generally anorectic (Asarian & Geary 2002, Asarian & Geary 2006, Geary et al 2001), although it varies between different neuronal populations (Smith et al 2014, Xu et al 2011). Estrogens’ effects are mediated fundamentally by the nuclear estrogen receptors (ERs) 𝛼 and 𝛽, especially ER𝛼 (Geary et al 2001, Liang et al 2002, Roesch 2006, Santollo et al 2010). However, estrogen responsive G-coupled membrane receptors, like GPR30, mediate some of their functions (Qiu et al 2006, Roepke et al 2010). In fact, some studies have found that the different receptors could be acting in combination (Filardo et al 2000).
The neuroprotective effects of steroids and neurosteroids have been widely demonstrated (Brotfain et al 2016, Day et al 2013, De Nicola et al 2013, Gold &
Voskuhl 2009, Rahmani et al 2016, Sarkaki et al 2011, Vegeto et al 2008, Wise 2003).
Aromatase, the enzyme that converts testosterone into estradiol, shows increased expression in reactive astrocytes following brain injury (Azcoitia et al 2003, García- Segura et al 1999, Saldanha et al 2009). Neuroprotection by sex steroids involves both astrocytes and microglia (Johann & Beyer 2013), as they decrease microglia reactivity (Cerciat et al 2010, Kipp et al 2007, Lewis et al 2008) and production of proinflammatory factors by astrocytes (Bruce-Keller et al 2000, Dimayuga et al 2005, Drew & Chavis 2000, Vegeto et al 2006).
3. Inflammation/gliosis in metabolism
Inflammation should not always be considered pathological, but as a protective response after an insult, necessary for tissue repair and control of homeostasis (Calder et al 2009). It is also a characteristic of obesity. However, obesity-linked inflammation is different from classical inflammation, as it lacks the rapid immune response and the
29
typical signs of heat, swelling, pain and redness. In contrast, inflammation associated to obesity is chronic and considered of low grade, in comparison to classical inflammatory responses in injury or infection (Gregor & Hotamisligil 2011, Medzhitov 2008).
Obese individuals, at least in adulthood, have high levels of circulating cytokines, such as TNF-𝛼, IL-1𝛽 and IL-6 (Berg & Scherer 2005, Hotamisligil et al 1995, Hotamisligil et al 1993, Shoelson et al 2006). This increase in cytokine levels is marked in adipose tissue (Fried et al 1998) and is combined with adipocyte hypertrophy, but not adipocyte proliferation (Weisberg et al 2003). This adipocyte hypertrophy is related to T cell and macrophage infiltration into the tissue (Minihane et al 2015). Interestingly, it has been found that visceral adipose tissue is more important in the inflammatory process than subcutaneous adipose tissue (Weiss 2007).
Other peripheral tissues also involved in obesity-related inflammation are the muscle (Saghizadeh et al 1996), liver (Cai et al 2005) and pancreas (Nicol et al 2013).
Importantly, inflammation associated to obesity is not only systemic, but it also occurs in the brain, and especially, the hypothalamus (Thaler et al 2012b). Besides adipose- borne cytokines that cross the BBB (Banks et al 1995, Benveniste 1998), cytokines can also be produced directly in the brain (Benveniste 1998). The activation and proliferation of astrocytes and microglia in response to an injury is known as gliosis, or reactive gliosis (Johns 2014), and frequently concurs with inflammation in the CNS.
Brain inflammation can be induced by an excess of nutrients (Aljada et al 2004, Cani et al 2007, Erridge et al 2007), especially long-chain saturated fatty acids, as they accumulate in the hypothalamus after HFD intake (Posey et al 2009). These nutrients and other inflammatory signals activate the toll-like receptor (TLR) 4 pathway. TLR4 activates JNK, IKKβ and other kinases, which interfere with the insulin-signaling cascade by phosphorylating IRS-1 (Boura-Halfon & Zick 2009, Tanti & Jager 2009) and stimulate cytokine production (Gorina et al 2011). This implies that HFD can have deleterious effects by itself; that is, these inflammatory effects are not solely related to excess body weight (Clegg et al 2011, Doerner et al 2016, Gao et al 2014). For example, hypothalamic inflammation/gliosis is a rapid response to HFD intake, occurring even before a gain of weight or any sign of adipose tissue inflammation can be observed (Thaler et al 2012b). This may be related to the cell stress that saturated fatty acids cause (Diaz et al 2015, Němcová-Fürstová et al 2013). However, the quick inflammatory reaction may be protective, as some studies have suggested (Buckman et al 2014). Thus, prolonged inflammation could possibly be involved in the perpetuation
30
of obesity and the development of some of its secondary complications (De Souza et al 2005, Douglass et al 2017, Milanski et al 2009). Other mechanisms, such as endoplasmic reticulum stress, have been found to be involved in hypothalamic inflammation (Hotamisligil 2010, Purkayastha et al 2011, Zhang et al 2008). Also, endoplasmic reticulum stress can induce autophagy (Butler & Bahr 2006, Yorimitsu et al 2006) which is related with the inflammatory pathways (Gan et al 2017, Muriach et al 2014).
4. Sex and age differences in response to HFD intake
During many years, most studies have been performed using only male individuals, alleging difficulties in the control of the estrous cycle in rodents and homogenization of the initial conditions. However, as some studies in females demonstrated important differences in the results compared to those obtained in males, the scientific community began to further explore such differences. In a similar way, most studies of obesity and dietary changes in rodents have used young adults as a standard. But, when performing similar experiments with pups, pubescents, middle aged or older adults, the results have often changed significantly, and sex differences have not always been reported to be the same.
4.1. Sex differences in young adults
High fat diet (HFD) intake causes weight gain at all ages, from prepubescents to adults (Grove et al 2010, Mela et al 2012b, Morselli et al 2014, Oliveira et al 2015, Venancio et al 2017, Williams et al 2014a). However, there is a sexual dimorphism in the effects of dietary habits. Young female adult rodents are usually less susceptible than males to weight gain and its secondary complications (Argente-Arizón et al 2016, Estrany et al 2013, Estrany et al 2011, Grove et al 2010, Hwang et al 2010, Mela et al 2012b, Morselli et al 2014, Oliveira et al 2015, Sánchez-Garrido et al 2013). In evolutionary terms, male mammals are reported to follow a strategy of increasing energy intake in order to raise fat stores (Shi et al 2009, Wade 1972), while females are more prone to preserve body fat by reducing energy expenditure (Shi et al 2009). These different strategies may be related to the differences in the hypothalamic melanocortin system (Mauvais-Jarvis 2015). That is, young male mice have fewer POMC neurons
31
and lower POMC expression than females (Nohara et al 2011) and are more sensitive to insulin, while females are more leptin sensitive (Clegg et al 2003).
Another difference between the sexes lies in the use of nutrients as energy sources. In resting conditions, women convert free fatty acids (FFA) to triglycerides (TG), thus storing fat, whereas FFA undergo higher levels of oxidation in men (Uranga et al 2005). On the other hand, when more energy is needed, women tend to use lipids and men tend to use carbohydrates (Carter et al 2001, Henderson 2014, Horton et al 1995).
Finally, one of the best-known differences between the body composition of men and women is the distribution of adipose tissue: women tend to store fat in subcutaneous pads, for long-term storage, and men tend to accumulate it in visceral pads, which are more metabolically active (Enzi et al 1986, Karastergiou et al 2012, Palmer & Clegg 2015, Vague 1947). Moreover, middle-age women usually present higher levels of brown adipose tissue (Cypess et al 2009, Rodrı́guez-Cuenca et al 2002).
In addition, the fat-borne hormones adiponectin and leptin have higher circulating levels in postpubertal women than in men (Considine et al 1996, Nishizawa et al 2002).
4.2. Sex differences at early ages
Although many of these sex differences are attributable to the effect of estrogens (Chowen et al 2018, Mauvais-Jarvis 2015), some studies have also found sex differences during the prepubertal period (Argente-Arizón et al 2016, Boukouvalas et al 2008, Krolow et al 2013), implying that sex steroids might not be the lone cause of this dimorphism. The effects of overweight, obesity and its secondary complications in children and immature animals are not directly comparable to what happens in adults (Martos-Moreno et al 2013, Pietrobelli et al 2008). These young individuals have a greater ability for tissue expansion, including adipose tissue, which could account for some of these age differences. For example, neonatal overnutrition results in increased weight gain in prepubertal male and female rats, but without showing any hypothalamic inflammation/gliosis, and this increased weight gain reappears only in adult males where signs of hypothalamic inflammation/gliosis are observed (Argente-Arizón et al 2018).
The effects of HFD intake on metabolism also differ between young adults and pubertal individuals. Puberty is defined as a critical phase of genital maturation and
32
development of secondary sex traits enabling sexual reproduction. It involves rapid growth, changes in body composition, fat redistribution and hormonal variations (Loomba-Albrecht & Styne 2009, Rogol 2010). Various studies have observed that early nutritional challenges induce an increase in body weight that then decreases or becomes non-existent during the pubertal period, only to reappear in adulthood (Argente-Arizón et al 2018, Habbout et al 2012, Mela et al 2012b, Stefanidis & Spencer 2012). These long-term effects of early nutritional and hormonal manipulations often do not become apparent until mid to late adulthood (Argente-Arizón et al 2018, Li et al 2013), suggesting that, during puberty and early adulthood, individuals have different susceptibilities to dietary or metabolic modifications. This raises the question as to what the responses to these kinds of challenges are when they occur specifically around pubertal onset. Secondary complications may also be affected by the timing of dietary challenges as, for example, HFD has a higher effect promoting mammary gland tumors when the intake takes places during puberty compared to adulthood (Zhu et al 2016).
4.3. Sex differences at old ages
There are multiple studies showing that females are more resistant to HFD- induced obesity and its secondary complications (Palmer & Clegg 2015, Pucci et al 2017). However, this statement has been recently challenged by other studies finding that middle-aged and old male and female mice exhibit the opposite response than their younger counterparts, with females being equally or more susceptible to the effects of poor dietary habits (Nishikawa et al 2007, Salinero et al 2018). The inversion of sex differences with age may be related to modifications in estrogen levels, which are high in young females, protecting them from the deleterious effects of obesity and metabolic challenges (Musatov et al 2007, Xu et al 2011), but diminish with age, with the consequent disappearance of their benefits (Leeners et al 2017, Nelson & Bulun 2001).
Indeed, a similar change occurs in humans, with post-menopausal women being more prone to obesity and metabolic syndrome (Kapoor et al 2017). The effects of testosterone on metabolic health and body composition have also been reported (Harada 2018, Kelly & Jones 2013) and circulating levels of this sex steroid may also decline with aging (Blaya et al 2017), but this modification is not a severe as that seen in women.
33
5. Interaction between metabolism, age and neurodegenerative diseases
It has been reported that obesity increases the risk of dementia (Emmerzaal et al 2015) and poor dietary habits, which can lead to neuroinflammation, can also produce cognitive dysfunction (Davidson et al 2013, Granholm et al 2008, Kanoski & Davidson 2011). Moreover, the prevalence of neurodegenerative diseases has been associated to obesity. The decrease of sex steroids that comes with aging has also been also highlighted as a factor raising the risk of cognitive deterioration (Manly et al 2000, Moffat et al 2004, Paoletti et al 2004, Pike et al 2009).
The most common cause of dementia is Alzheimer’s Disease, a neurodegenerative process that was first described more than a hundred years ago (Alzheimer 1907) and remains one of the major health challenges as life span continues to increase, resulting in a higher prevalence of this disease. Unfortunately, an effective treatment remains elusive. AD implies a progressive cognitive decline, with memory loss and impaired reasoning. Its more characteristic features are high cortical concentrations of neurofibrillary tangles, composed of hyper-phosphorylated tau protein, and neuritic plaques of amyloid-β peptide (Aβ) (Braak & Braak 1997). The most accepted hypothesis for AD’s pathogenesis is that soluble Aβ oligomers affect synaptic functioning and unleash inflammation pathways, an imbalance in central calcium concentrations and oxidative stress, leading to cognitive dysfunction (Selkoe 2000). The Aβ peptide is a cleavage product of the amyloid protein precursor (APP) (LaFerla et al 2007). Mutations of the APP gene or in genes encoding for presenilin 1 (PS1), a protein involved in APP cleavage (Hardy & Selkoe 2002), lead to an increase of Aβ levels and could account for the inheritable, early-onset form of the disease (Lane et al 2018, Pimplikar 2009). However, most cases of AD are of late onset, in which the impairment of Aβ clearance is possibly the cause of its accumulation (Mawuenyega et al 2010). Nevertheless, there are genetic factors involved in late-onset AD, such as mutations in APOE, a gene that encodes the apolipoprotein E (ApoE), a lipid carrier (Bu 2009, Farrer et al 1997, Liu et al 2013). ApoE malfunctioning may be related to oxidative stress, synaptic loss and formation of amyloid plaques (Bell et al 2007, Hashimoto et al 2012, Miyata & Smith 1996, Sen et al 2012). The insulin signaling pathway has been shown to be affected in some AD patients (Messier & Teutenberg 2005) and high circulating glucose levels can be involved in the defective clearance of
34
A𝛽 and the formation of neurofibrillary tangles (Querfurth & LaFerla 2010). These observations indicate a possible link between AD and energy metabolism.
The complex etiology of AD makes it difficult to establish animal models that parallel the causes and progression of the disease in humans. Although AD does not appear in rodents, there are some acceptable transgenic mice models that are widely used in the study of this disease, presenting mutations in genes APP, APP and PS1, APP and APOE and APP with insulin desensitization (Esquerda-Canals et al 2017). These models have been employed in attempt to further our understanding of this devastating disease.
The incidence of AD is higher in women than in men, probably due to our different levels of sex steroid hormones (Li & Singh 2014, Moser & Pike 2016), and some observations are similar in genetic animal models. Indeed, Aβ depositions appear earlier in female than in male transgenic AD mice, as well as cognitive deficits (Henderson & Buckwalter 1994, Yang et al 2018). Understanding the interaction between aging, sex and dietary habits in neurodegenerative processes could improve the chances of finding effective treatments for them.
35
HYPOTHESIS
36
37
Hypothesis:
There is an intricate interaction between genetics, sex and
age in the response to poor dietary habits and this underlies the
differences observed in weight gain and the development of
secondary complications related to excess body weight.
38
39
OBJECTIVES
40
41
The specific objectives of this thesis were:
Objective 1. To analyze the response of peripubertal animals to a short- term exposure to a HFD and determine if males and females respond differently.
Objective 2. To study the interaction between long-term HFD exposure, sex and genetic over-expression of beta-amyloid on weight gain and metabolic complications.
Objective 3. To determine the response of hypothalamic astrocytes to fatty acids, sex steroids and beta-amyloid.
42
43
MATERIAL AND METHODS
44
45 1. Experimental models
1.1. In vivo studies
All of the studies were designed according to the European Community Council Directive (86/609/EEC; 2010/63/UE) and NIH guidelines for animal care and complied with the Royal Decree 53/2013 pertaining to the protection of experimental animals.
The studies and use of animals were approved by the animal care and use committee of the Cajal Institute (Comité de Experimentación Animal del Instituto Cajal) and the Consejeria del Medio Ambiente y Territorio (Comunidad de Madrid, Ref. PROEX 200/14).
1.1.1. Response of peripubertal animals to short-term exposure to a HFD
a) Experimental design
Mice from C57BL/6Jola strain (5 weeks-old) were purchased from Harlan Interfauna Ibérica S.A., (Barcelona, Spain). Four animals of the same sex were placed in each cage and allowed to acclimate for at least one week before the start of the experiment. During this time, they were fed a normal chow diet and water ad libitum.
All mice were maintained at a constant temperature (21 ± 1ºC) and humidity (50 ± 1%) in a 12-hour light-dark cycle.
At post-natal day (PND) 42, day one of the study, animals were weighed and mice of each sex were randomly separated in two groups receiving either a low-fat diet (LFD; 3.76 Kcal/g, 10.2% fat; LabDiet, Sodispan Research SL, Madrid, Spain) or high- fat diet (HFD; 5.1 Kcal/g, 61.6% fat; LabDiet). The composition of each diet is shown in Table 1. Food intake during 24 hours was measured on 3 different days throughout the week of the experiment by placing a known amount of food in the cage and weighing the remaining food the following day. Mean energy intake is reported as total kilocalories consumed and the number of kilocalories consumed/ gram bodyweight.
b) Sacrifice and sample collection
On day 7, after weighing the remaining food the animals were subjected to a 12 hour fast and killed on the morning of day 8 by decapitation (Figure 3). Their hypothalamus, hippocampus, cerebellum, liver and subcutaneous and visceral fat pads were extracted. The brains and fat pads were weighed, and the fat pads frozen in dry ice.
46
Dissection of the brains was performed rapidly and on a cold surface before freezing in dry ice. Trunk blood was collected, allowed to clot, centrifuged (3000 rpm during 10 min at 4 °C) and the serum removed. All tissues were stored at −80 °C until processed. Glycemia was measured at the moment of sacrifice using a glucometer (Optium Xceed, Abbott Diabetes Care, Inc. CA, USA).
Figure 3. Experimental design for Objective 1.
1.1.2. Interaction between long-term HFD exposure, sex and genetic over- expression of beta-amyloid on weight gain and metabolic complications.
a) Experimental design
Transgenic knock-in heterozygous mice that over-express amyloid precursor protein (TgAPP) mice were employed for this study. These mice were created on a C57BL/6 background by heterozygous breeding of mice expressing the human APP long isoform with two mutations [Lys 670- Asn and Met 671-Leu (Swedish mutation)]
under transcriptional control of the hamster prion promoter (Hsiao et al 1996). Wild type (WT) littermates were used as controls. Four animals of the same sex, but of both genotypes (WT or TgAPP) were placed in each cage and allowed to acclimate for at least one week before the start of the experiment.
During this time, they were fed a normal chow diet and water ad libitum. All mice were maintained at a constant temperature (23 ± 2ºC) and humidity (55 ± 1%) in a
47
12-hour light-dark cycle. Fifty-seven mice (11 WT males, 18 APP males, 10 WT females and 18 APP females) 7 months old were randomly separated in two groups (5 WT LFD males, 6 WT HFD males, 9 APP LFD males, 9 APP HFD males, 5 WT LFD females, 5 WT HFD females, 9 APP LFD females and 9 APP HFD females) receiving either a LFD ( 3.76 Kcal/g, 10.2% fat; LabDiet) or HFD (5.1 Kcal/g, 61.6% fat;
LabDiet). See table 1 for composition of the diets. This resulted in 8 groups of mice:
male and female WTLF, WTHF, APPLF and APPHF. The mice were maintained on these diets for 18 weeks and body weight was measured weekly.
LFD HFD
Carbohydrates: 67.4 % 25.9 %
Sugars 33.13 % 8.85 %
Fats: 4.3 % 34.9 %
Cholesterol 18 ppm 301 ppm
Saturated FA 1.14 % 13.68 %
Monounsaturated FA 1.3 % 14 %
Polyunsaturated FA 1.59 % 5.15 %
Proteins 16.9 % 23.1 %
Fiber 4.7 % 6.5 %
Table 1. Diet composition in percentage of weight (except cholesterol, which is in parts per million).
b) Sacrifice and sample collection
The mice were subjected to a 12 hour fast before sacrifice. They were weighed and sacrificed by decapitation and their hypothalamus, hippocampus, cerebellum, pituitary and subcutaneous and visceral fat pads extracted. The fat pads were weighed and frozen in dry ice. Dissection of the brains was performed rapidly on a cold surface before freezing in dry ice. Trunk blood was collected, allowed to clot, centrifuged (3000 rpm during 10 min at 4 °C) and the serum removed. All tissues were stored at
−80 °C until processed. Glycemia was measured at the moment of sacrifice by using a glucometer (Optium Xceed, Abbott Diabetes Care, Inc. CA, USA). A summarized representation of the experimental design can be consulted in Figure 4.
48 Figure 4. Experimental design for Objective 2.
1.2. In vitro studies
1.2.1. Primary hypothalamic astrocyte cultures
Primary cultures of hypothalamic astrocytes were prepared from post-natal day two (PND2) Wistar rat pups. Both sexes were used, and male and female cells were cultured separately after determining the animal’s sex by anogenital distance. Pups were sacrificed by decapitation and the brain quickly removed and immediately placed on ice in Dulbecco’s modified Eagle’s medium: Nutrient mixture F-12 (DMEM/F12) supplemented with 1% penicillin/streptomycin and anti-mycotic (Ab/Am; Gibco).
Hypothalami were dissected and the meninges carefully removed. Hypothalami were then triturated in the same media as indicated previously. The tissue was dissociated by first using a P1000 pipette and then pulled Pasteur pipettes several times until the suspension was clear. The suspension was then centrifuged at 1000 rpm for 7 min. The resulting pellet was resuspended in DMEM/F12 enriched with 10% fetal bovine serum (FBS) plus 1% Ab/Am solution. Cells were then seeded in 75 cm2 culture flasks (BD Falcon) containing 9 mL of DMEM/F12 plus 10% FBS and 1% Ab/Am. The cells were incubated at 37 °C and 5% CO2. After three days, the flasks were washed twice with tempered phosphate buffered saline (PBS) and fresh media was added. Media was changed three times a week and after a total of 9-10 days of incubation, when the cells reached approximately 90% confluency, flasks were placed in a 37 °C shaking
49
incubator (SI-300, Jeoi Tech; Medline Scientific) at 280 rpm overnight. A summarized representation of this experimental design can be seen in Figure 5.
After shaking, flasks were washed twice with PBS to remove non-attached cells.
Then, 1 mL of trypsin (0.05% trypsin/EDTA solution; Biochrom AG) was added and the flasks hit together 10 times (in order to detach and harvest the astrocytes).
DMEM/F12 plus 10% FBS and 1% Ab/Am was added to the flasks to stop the action of the trypsin and resuspend the cells. The suspension was centrifuged for 5 min at 1150 rpm. After centrifugation, the supernatant was discarded and the pellet resuspended in DMEM/F12 plus 10% FBS and 1% Ab/Am. Astrocytes were seeded in 60 mm or 100 mm culture plates, that had been previously treated with poly-L-lysine hydrobromide (10 µg/ml; Sigma-Aldrich), at a density of 4.35 x 105 or 1 x 106 cells/plate, respectively.
Cells were then grown for 24 hours and then the media was changed to DMEM F-12 plus 1% Ab/Am (without FBS). Thus, the cells were serum starved for 24 h before the experimental treatments were added. The treatments were prepared with the same media (DMEM F-12 plus 1% Ab/Am, without FBS). In each experiment, treatments were done in triplicate; each experiment was repeated 3 to 4 times (N = 3 or 4).
Figure 5. Experimental design of primary cultures of hypothalamic astrocytes.
50 1.2.2. Treatments
a) Palmitic acid
All reagents were purchased from Sigma-Aldrich, Inc. (Saint Louis, MO, USA).
Palmitic acid (PA) supplemented medium was prepared according to previously published protocols (Huynh et al., 2014) and used to treat astrocytes. Stock solutions of PA (Sigma, P-5585), BSA (Sigma, A-9205) and L-carnitine (Sigma, C-0283) were prepared and sterilized by filtration with 0.45 µm filters and stored at -20°C until used.
The PA stock solution was prepared at 200 mM in ethanol and vortexed until the solution was clear. Fatty acid free bovine serum albumin (BSA) was added to ensure lipid solubility in the aqueous solution The BSA stock solution was prepared at 30%
and L-carnitine at 200 mM; both substances were dissolved in sterile H2O. Palmitic acid was then conjugated with fatty acid-free BSA to act as a carrier and ensure lipid solubility in the aqueous solution. Fatty acid-albumin solutions were diluted in DMEM F-12 plus 1% Ab/Am (without FBS) to achieve the desired final fatty acid concentration. The L-carnitine (1 mM) was added to the final fatty acid solution before being added to cultures. Control plates received an equivalent amount of vehicle solution. Treatments had a duration of 24 hours.
b) Estradiol
A stock solution of 1 mg/mL of 17β-estradiol (Sigma, E-8875) was prepared in ethanol. Then, a 10-9 M working solution was made in DMEM F-12 plus 1% Ab/Am (without FBS) medium for treatment of cultured astrocytes.
When 17β-estradiol was used in combination with PA, astrocytes were pre- treated with 10-9 M β-estradiol for 3 h previous to PA addition. Once all the treatments were added, they were incubated for 24 hours.
c) Testosterone
Testosterone (Sigma, T-1500) was purchased from Sigma-Aldrich. A stock solution of 0.1 mg/mL was prepared in ethanol. Then, 10-8 M, 10-9 M and 10-10 M working solutions were made in DMEM F-12 plus 1% Ab/Am (without FBS) medium for treatment of cultures astrocytes. Testosterone was added to the cultures at the same time as PA and vehicle (controls) and all the plates were incubated with the treatments for 24 hours.
51 d) Amyloid-𝜷
Amyloid-𝛽 protein 1-40 (SC875) was purchased from PolyPeptide (Strasbourg, France). A stock solution of 1 mg/mL was prepared in PB 0.05 M. Before treatment, the volume of stock solution needed was incubated at 37ºC for 24 hours, with occasional vortexing. This process allows the peptide to aggregate and therefore induce neurotoxicity in vitro (Iversen et al 1995). Then, 1, 5 and 20 𝜇g/mL working solutions were made in DMEM F-12 plus 1% Ab/Am (without FBS) medium for 24 hours treatment of cultured astrocytes.
2. Techniques and protocols 2.1. Serum analysis
Circulating leptin, insulin, interleukin (IL) 6 and tumor necrosis factor α (TNFα) for both Objective 1 and Objective 2 and monocyte chemoattractant protein (MCP)-1 and total plasminogen activator inhibitor (PAI)-1 for Objective 2, were determined by multiplexed immunoassay according to the manufacturer’s instructions (Millipore, Billerica, MA) in a Bio-Plex suspension array system 200 (Bio-Rad Laboratories, Hercules, CA, USA). Mean fluorescence intensity was analyzed by using Bio-Plex Manager Software 4.1. All samples were run in duplicate and within the same assay for all analyses. The minimum detectable concentrations of IL-6, insulin, leptin, MCO-1, PAI-1 Total and TNF𝛼 were 2.3, 13.0, 4.2, 4.9, 4.0 and 5.3 pg/ml, respectively. The intra-assay coefficients of variation were 5 % for all analytes except for TNF𝛼, which was 4 %. The inter-assay coefficients of variation of IL-6, insulin, leptin, MCO-1, PAI- 1 Total and TNF𝛼 were 11, 11, 10, 10, 15 and 20 %, respectively.
Circulating triglycerides (Spinreact S.A., Sant Esteve de Bas, Spain) and non- esterified fatty acids (NEFA; Wako, Neuss, Germany) were measured using enzymatic colorimetric kits according to the manufacturers’ instructions.
2.2. RNA and protein isolation
Total mRNA was isolated from non-adipose tissues or astrocyte cultures by using an RNeasy® Plus Mini Kit (Qiagen, Hilden, Germany). For adipose tissue, a RNeasy® Lipid Tissue Mini Kit (Qiagen) was used. Protein was isolated from the same
52
tissues by collecting the first elution from the RNeasy® Mini Spin columns and diluting 1:5 with cold acetone. It was then stored at -20 ºC for at least 30 minutes before centrifuging (3000 rpm during 10 min at 25 ºC), after which the acetone was removed and the pellet resuspended in 100 µl of 3-[(3-cholamidopropyl) dimethylammonio]
propanesulfonate (CHAPS) buffer (7 mM urea, 2 M thiourea, 4% [w/v] CHAPS, and 0.5% [v/v] 1 M Tris; pH 8.8). Protein samples were stored at -20 ºC until protein quantification by the method of Bradford (Bio-Rad Laboratories). Determination of mRNA purity and concentration was performed by using a Nanodrop (Thermo Scientific, Washington, DE, USA) and samples were stored at -80 ºC until analyzed.
2.3. RT-qPCR
Using 1 µg of each RNA sample, cDNA was synthesized with a high-capacity cDNA RT kit (Applied Biosystems, Foster City, CA) and stored at -20 ºC. Quantitative RT-PCR was performed by using TaqMan Universal PCR Master Mix (Applied biosystems) and TaqMan Gene Expression Assay-on-demand kits to analyze neuropeptides and receptors involved in metabolic control (Table 2). All samples were run in duplicate. Various housekeeping genes were tested and those that did not vary between experimental groups were chosen to normalize the data (Table 2). The ∆∆CT method was used to determine relative expression levels and for statistical analysis. All data are expressed as % control group.
53
Gene Reference Animal
GFAP Mm01253033_g1 Mouse
Rn00566603_m1 Rat
DDIT-3 Mm01135937_g1 Mouse
Rn00492098_g1 Rat
IL-6 Mm00446190_m1 Mouse
Rn01410330_m1 Rat
IL-1β Mm01336189_m1 Mouse
TNFα Mm00443260_g1 Mouse
POMC Mm00435874_m1 Mouse
NPY Mm03048253_m1 Mouse
AgRP Mm00475829_g1 Mouse
Leptin Mm00434759_m1 Mouse
LepR Mm00440181_m1 Mouse
InsR Mm01211875_m1 Mouse
IκBKβ Mm01222247_m1 Mouse
StAR Rn00580695_m1 Rat
TSPO Rn00560892_m1 Rat
Aromatase Rn00567222_m1 Rat
18S Mm03928990_g1 Mouse
Rn01428915_g1 Rat
Pgk1 Mm00435617_m1 Mouse
Rn00821429_g1 Rat
Rpl13a Rn00821946_g1 Rat
Ppia Mm02342430_g1 Mouse
Actin beta Mm00607939_s1 Mouse
Rn00667869_m1 Rat
GAPDH Mm99999915_g1 Mouse
Rn99999916_s1 Rat
Table 2. List of genes analyzed by qRT-PCR.
54 2.4. Western blotting
For Western blotting, equal quantities of protein of each sample were resolved on an SDS- acrylamide gel under denaturing conditions (see Table 2 for protein quantities and acrylamide concentrations). Proteins were then transferred to a polyvinylidine difluoride membrane (Bio-Rad, Hercules, CA, USA). Transfer efficiency was determined by Ponceau red dyeing. Membranes were blocked with Tris-buffered saline (TBS) containing 5% (w/v) non-fat milk or bovine serum albumin (phosphorylated proteins) and incubated with the appropriate primary antibody and concentration overnight at 4 ºC under agitation. The antibodies and their concentrations used are listed in Table 3.
The following day after washing, membranes were incubated with the secondary antibody conjugated with peroxidase (Pierce Biotechnology, 1:1000 or 1:2000).
Peroxidase activity was visualized by using chemiluminescence and quantified by densitometry using an Image-Quant LAS4000 mini TL Software (GE Healthcare Europe GmbH, Spain). All blots were rehybridized with actin to adjust for loading and then normalized to % control group values on each gel.
55
Antigen Host Company Dilution
Actin, pan Ab-
5 Mouse Thermo Fisher
Scientific 1:1000 GAPDH Rabbit AnaSpec 1:1000 GFAP Mouse Sigma-Aldrich 1:1000 Vimentin Rabbit Sigma-Aldrich 1:1000 IBA-1 Rabbit Wako 1:1000
JNK (F-3) Mouse Santa Cruz
Biotechnology 1:500
pJNK Mouse Promega 1:500
SOD Rabbit Sigma-Aldrich 1:500
Hsp-70 Rabbit Enzo 1:1000
pNF𝜿B Rabbit Cell Signalling 1:500
NF𝜿B Mouse Thermo Fisher
Scientific 1:500 pAKT Mouse Cell Signalling 1:500
AKT Goat Santa Cruz
Biotechnology 1:500 pI𝜿B𝜶 Mouse Cell Signalling 1:500 pIRS1 Mouse Cell Signalling 1:500 IRS1 Rabbit Sigma Aldrich 1:500
Table 3. List of antibodies used in Western Blots.
2.5. Estimation of the number of cells
To determine cell viability in response to amyloid-𝛽 treatment in primary hypothalamic astrocyte cultures we used the crystal violet dye elution method. The dye binds to the nuclei of cells and the amount of dye is quantified by solubilizing it and using a spectrophotometer. The absorbance is proportional to the number of cells.
